Computer system and method for detecting and processing changes in data

• US 8,620,716 B2
• Filed: 12/05/2003
• Issued: 12/31/2013
• Est. Priority Date: 12/06/2002
• Status: Active Grant

• Current Assignee: SAP America Incorporated (SAP SE)
• Original Assignee: SeeWhy, Inc. (SAP SE)
• Inventors: Thomas, Paul, Nicholls, Charles Martin
• Primary Examiner(s): Fleischer, Mark A
• Application Number:
  US10/538,225
• Publication Number:
  US 20060167704A1
• Time in Patent Office: 3,679 Days
• Field of Search: 705/7.29
• US Class Current: 705/7.29
• CPC Class Codes: G06Q 10/06393 Score-carding, benchmarking or key performance indicator [KPI] analysis
• CPC Class Codes: G06Q 10/10 Office automation; Time management

Abstract

The system enables business people to understand the impact of business transactions, changes and events in real-time using advanced rules and analytics to filter, categorize and interpret the significance of streams of real-time information. Most business performance analysis today is done manually and this process is a time consuming and skilled task leading to a time delay in producing the analysis. This time lag between the transaction or event happening and being able to take action on the analysis is measured in weeks or months at many companies. By blending real-time information with historical data and performance goals, this system enables business users to assess business events and collaborate within teams to drive optimal business performance. Using forecasting techniques enables business managers to predict the likelihood of achieving a particular goal without relying on manual analysis by a skilled analyst. The system automatically updates the forecast based on real-time changing data, enabling the business manager to have an up to the minute and statistically valid projection of future business performance.

Claims

1. A computer system for managing data, the system comprising:

a memory storing historical data comprising continuously updated values for each of a plurality of metrics at a plurality of points in time; and

at least one computer configured to;

from one or more external sources, receive a continuous stream of event data for at least one of the metrics, wherein the event data is transmitted to and received by the at least one computer independent of a request from the computer,

in response to each received event data, calculate with the computer an actual value for at least one metric based at least in part on the event data, wherein the actual value is calculated based on a first sliding time window, and represents a current state of the at least one metric,

in response to each received event data, automatically calculate with the computer a computed value for the at least one metric, wherein the computed value is calculated based on a second sliding time window, and represents an expected state of the at least one metric at a time corresponding to a time associated with the actual value,

compare the computed value with the actual value, and

determine whether to generate an alert or action based on the comparison of the computed value and the actual value,

wherein the computed value is calculated in response to each received event data, and the computed value is calculated based at least in part on the second sliding time window being defined relative to and inclusive of a time the received event data prompting the automatic calculation was received and further defined by one or more of;

a time of a most recently received event data of the continuous stream of event data,

a last hour of the continuous stream of event data,

a last number of hours, x, of the continuous stream of event data, where the number of hours, x, has been stored in the computer system,

another metric that has been stored in the system,

a stored last number of a particular day of the week, thereby including data only from those particular days of each week on which the continuous stream of event data was received,

a stored last number of a particular day of the month, thereby including data only from those particular days of each month on which the continuous stream of event data was received, and

a stored last number of a particular day of the year, thereby including data only from those particular days of each year on which the continuous stream of event data was received.

2. The system according to claim 1, wherein the at least one computer is arranged to store one or more rules for operating on at least one of the received event data, the actual value, the computed value, and the comparison of the actual value and the computed value and wherein a metric is part of the rules and the metric is defined by an operator.

3. The system according to claim 1, wherein the at least one computer is arranged to store one or more alert definitions for causing a signal to be sent according to the rules, wherein criteria for the rules is determined either by a fixed value, or by a dynamic metric in the system.

4. The system according to claim 3, wherein the at least one computer is arranged to send a message to a terminal selected based on the alert.

5. The method of claim 1, wherein the metrics are defined by an operator of the computer system.

6. The method of claim 1, further comprising reporting the comparison of the computed value and the actual value.

7. The method of claim 1, further comprising:

notifying one or more users of the computer system of the alert or action;

storing status information to reflect the priority of the alert or action;

storing information related to the alert or action;

storing status information after the action has been taken; and

correlating previous actions taken with performance and achievement of goals.

9. The method of claim 1, further comprising:

automatically generating associating a dimension of the metric;

automatically generating associating additional individual level metrics, without operator involvement where the individual level dimension has not been explicitly defined previously or by an operator;

automatically interpreting the new metric over time; and

automatically calculating another computed value.

8. A computer system for managing data relating to the performance of an enterprise, the system comprising:

a memory storing historical business data comprising values of each of a plurality of metrics at a plurality of points in time; and

at least one computer configured to;

from one or more external sources, receive a continuous stream of event data for at least one of the metrics, wherein the event data is transmitted to the at least one computer independent of a request from the computer,

in response to each received event data, calculate with the computer an actual value for at least one metric based at least in part on the event data, wherein the actual value is calculated based on a first sliding time window, and represents a current state of the at least one metric,

in response to each received event data, calculate with the computer a computed value for the at least one metric based on the historical business data. wherein the computed value is calculated based on a second sliding time window and represents an expected state of the at least one metric at a time corresponding to a time associated with the actual value,

wherein the computed value is calculated in response to each received event data, and the computed value is calculated based at least in part on the second sliding time window being defined relative to and inclusive of a time the received event data prompting the automatic calculation was received and further defined by one or more of;

a time of a most recently received event data of the continuous stream of event data,

a last hour of the continuous stream of event data,

a last number of hours, x, of the continuous stream of event data, where the number of hours, x, has been stored in the computer system,

another metric that has been stored in the system,

a stored last number of a particular day of the week, thereby including data only from those particular days of each week on which the continuous stream of event data was received,

a stored last number of a particular day of the month, thereby including data only from those particular days of each month of which the continuous stream of event data was received, and

a stored last number of a particular day of the year, thereby including data only from those particular days of each year on which the continuous stream of event data was received.

10. A computer system for managing data, the system comprising:

a memory storing historical data comprising continuously updated values for each of a plurality of metrics at a plurality of points in time; and

at least one computer configured to;

from one or more external sources, receive a continuous stream of retail sales event data, wherein the retail sales event data is transmitted to and received by the computer independent of a request from the computer,

in response to each received retail sales event data, calculate with the computer an actual retail sales volume value for a retail sales volume metric based at least in part on the retail sales event data, wherein the actual retail sales volume value is calculated based on a first sliding time window, and represents a current state of the retail sales volume metric,

in response to each received retail sales event data, automatically calculate with the computer a computed value for the retail sales volume, wherein the computed value is calculated based on a second sliding time window, and represents an expected state of the current retail sales volume at a time corresponding to a time associated with the actual retail sales volume value,

compare the computed value with the actual value, and

determine whether to restock or to change the price of a product based on the comparison of the computed value and the actual value,

wherein the computed value is calculated in response to each received event data, and the computed value is calculated based at least in part on the second sliding time window being defined relative to and inclusive of a time the received event data prompting the automatic calculation was received and further defined by one or more of;

a time of a most recently received event data,

a last hour of the continuous stream of retail sales event data,

a last number of hours, x, of the continuous stream of retail sales event data, where the number of hours, x, has been stored in the computer system,

another metric that has been stored in the system,

a stored last number of a particular day of the week, thereby including data only from those particular days of each week on which the continuous stream of event data was received,

a stored last number of a particular day of the month, thereby including data only from those particular days of each month on which the continuous stream of event data was received, and

a stored last number of a particular day of the year, thereby including data only from those particular days of each year on which the continuous stream of event data was received.

11. The computer system of claim 10, wherein the retail sales event data comprises a retail sale of the product.

12. The computer system of claim 10, wherein the retail sales volume metric comprises a sales rate of the product, and the actual retail sales volume value comprises a sales rate of the product over the sliding time window.